XAI-Powered Hybrid Model for Real-Time Financial Fraud Detection

P. Devika, G. Mathu Kumar, P. Nagul Kumar, S. Naveen Prabhu

2025

Abstract

With the rising complexity of false exercises, the discovery of extortion in money related trades has gotten to be a major challenge. Conventional machine learning models frequently act as dark boxes, making it greatly troublesome to clarify the choices they make. To address this, this extension proposes a cross breed extortion location that combines Irregular Timberland and XGBoost models with XAI procedures such as SHAP to encourage the creation of logical AI touchy to the thinking behind extortion expectations. The framework is prepared on a Kaggle dataset of budgetary exchanges, with a Java backend and HTML, CSS, and JavaScript frontend. The cross-breed show utilizes weighted averaging to coordinate both calculations, making extortion location strong and dependable. In terms of the XAI viewpoint, the system gives human-readable clarifications, such as highlighting bizarre exchange sums, login irregularities, and suspicious geographic designs. This system addresses a few real-time challenges related to extortion location, explainability, and ill-disposed strength, eventually displaying a clean and palatable arrangement for the budgetary division to improve their extortion avoidance procedures. The proposed system guarantees that any yields given are in full compliance with straightforwardness, operational guidelines, and administrative rules, reestablishing certainty in AI's capacity to identify budgetary wrongdoings.

Paper Citation

in Harvard Style

Devika P., Kumar G., Kumar P. and Prabhu S. (2025). XAI-Powered Hybrid Model for Real-Time Financial Fraud Detection. In Proceedings of the 1st International Conference on Research and Development in Information, Communication, and Computing Technologies - ICRDICCT`25; ISBN 978-989-758-777-1, SciTePress, pages 214-218. DOI: 10.5220/0013880300004919

in Bibtex Style

@conference{icrdicct`2525,
author={P. Devika and G. Kumar and P. Kumar and S. Prabhu},
title={XAI-Powered Hybrid Model for Real-Time Financial Fraud Detection},
booktitle={Proceedings of the 1st International Conference on Research and Development in Information, Communication, and Computing Technologies - ICRDICCT`25},
year={2025},
pages={214-218},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0013880300004919},
isbn={978-989-758-777-1},
}

in EndNote Style

TY - CONF

JO - Proceedings of the 1st International Conference on Research and Development in Information, Communication, and Computing Technologies - ICRDICCT`25
TI - XAI-Powered Hybrid Model for Real-Time Financial Fraud Detection
SN - 978-989-758-777-1
AU - Devika P.
AU - Kumar G.
AU - Kumar P.
AU - Prabhu S.
PY - 2025
SP - 214
EP - 218
DO - 10.5220/0013880300004919
PB - SciTePress